Do op amp low pass filters stop changes in DC voltages?

[JohnJohn20]

OK. A low-pass filter is a filter that passes low-frequency signals and attenuates (reduces the amplitude of) signals with frequencies higher than the cutoff frequency.

I understand the words but, how does a low pass filter affect an increase in the input voltage?

For example, what will happen to the output of a unity gain op amp low pass filter if the input voltage increases from 2V to 4 V over 0.5 seconds?

Do we treat this as an AC signal with a period of 2 seconds or 1 second or what?

What is the minimum cutoff frequency of this op amp filter that will allow this change in input voltage but stop any higher frequency noise?

FYI. I am trying to monitor a high impedance (~5MΩ) voltage for changes in voltage that take between 0.25 and 0.5 seconds to happen.

I have a sensitive detector but I am having a lot of problems with noise getting into the detector input. Any suggestions would be appreciated.

Would a two pole low pass unity gain op-amp filter circuit be the right way to approach this problem?

Thanks for reading.

 

[reidintransit]

What frequency is the noise? the number of poles will be determined by the ratio of the frequency you want to keep to the frequency you need to eliminate. You may just need a choke or a shunt capacitor.

 

[JohnJohn20]

I am not sure and don't have anything to measure it. I assumed it was 50Hz. I get the noise when I touch the input lead with my finger.

By choke do you mean wrapping the wires around a ferrite bead similar to that found on a laptop PSU cord or a usb cord? Or a stronger inductor?

I did try putting a 0.1uF and a 2.2uF capacitor across the voltage input but that made it made my ADC readings jump around more and didn't seem to quieten down the noise introduced with my finger touch.

 

[crutschow]

Yes, you treat the DC change as a low frequency AC signal. The LP filter frequency rollover point must be high enough to allow this change and still settle within the time period you require.

So we need to know:

How fast the new signal needs to settle to it's new value and to what accuracy?

What is the noise frequency range and amplitude and how much do you need to suppress it?

How are you reading the signal?

 

[reidintransit]

Couple of thoughts on touching live circuits:
1. Your body can have a static charge of electrons big enough to ruin electronics. Great rule of thumb is to keep your fat thumbs off sensitive electronics.
2. If you hold onto the line you are either acting like an antenna for the ambient power EMI and other signals directly into your circuit or a ground for the signal in your circuit.
Unless you are planning on having the design attached to a human, I don't see the point of touching the input.

 

[crutschow]

If it's 50Hz pickup then you could perhaps use a 50H notch filter.

Are you using shielded wires for the signal? If not, you need to.

 

[JohnJohn20]

OK crutschow. Thank you.

I am not sure but I believe the voltage change will happen over 0.25 to 0.5 seconds.

I don't know what the noise looks like. I am detecting it using an PIC ADC and sending the data to a computer which displays the number values. The PIC clock is oscillating at 1Mhz, it sends about results per minute and the injected noise just makes the numbers go haywire. Sorry I can't be more specific.

I did fiter the signal with a 20Hz 2 pole op amp LP filter which didn't seem to help much.

I will try retuning it to 1Hz and see if that helps.

 

[crutschow]

If you don't have an oscilloscope to look at the noise then it can be difficult to guess at its source and a solution to reduce it.